Editorial: Security and Dependability of Cloud Systems and Services - Part II

THIS is the second part of the Special Issue on Security and Dependability of Cloud Systems and Services, which we organized to solicit novel results in these important and closely related research areas. Indeed, security and dependability are increasingly important concerns for cloud systems and services, due to their spread in a large variety of application fields, including businessand mission-critical scenarios. This demands for innovative methodologies, algorithms and techniques for building services in which companies, organizations and citizens can trust and rely upon. The call attracted seventy-eight submissions from thirtyone countries; all of them underwent a thorough review process, which resulted in eleven selected manuscripts, with an acceptance rate of 14 percent. The eleven articles were split into two parts. The first part of the special issue focused on security aspects from the perspective of both providers and users. The six manuscripts addressed challenges in the areas of cloud trust management, secure cloud storage systems, and security-related service level agreements. This second part includes five papers which address relevant dependability and security issues concerning services in federations of cloud platforms, in mobile cloud computing, in data centers and in sensor clouds. Federations of cloud services and resources allow organizations to use a combination of different clouds—across different providers and tenants—to satisfy the needs of their applications, made up by a collection of diverse software components. Dependability and security are important concerns for such applications, but they are typically addressed separately. The first article, entitled “Cost Effective, Reliable and Secure Workflow Deployment over Federated Clouds” by Zhenyu Wen, Jacek Cala, Paul Watson, and Alexander Romanovsky, consider altogether issues of security, reliability and cost. The problem targeted is the deployment ofworkflow applications on federated clouds. The paper presents an algorithm which takes into account reliability, security and monetary cost of the deployment; the latter considers computing, storage and communication costs. The algorithm is shown to provide solutions which meet security requirements while providing reliability equivalent to two popular multi-criteria scheduling algorithms, at a cheaper cost. In the second article, entitled “Risk Assessment in a Sensor Cloud Framework Using Attack Graphs”, Amartya Sen and Sanjay Madria deal with security risk assessment in a sensor cloud. Sensor clouds are collections of heterogeneous wireless sensor networks (WSNs)—in general with different ownerships—whose sensing services are provided through a cloud platform. They may be subject to security threats both in the deployment areas, where sensors are not physically monitored for long periods or may even be hostile environments, and in the cloud integration environment. The authors propose a risk assessment framework for WSNs in a sensor cloud, based on attack graphs; the attacks may target confidentiality, integrity or availability. The framework supports security administrators in better comprehending the threats and take necessary actions, as well as to determine the efficiency of a security measure. It is validated by comparisonwith a variety of simulated attack scenarios. The cloud computing paradigm has become popular also for mobile applications: indeed, with a market value currently estimated in several billionUS dollars per year,Mobile Cloud Computing (MCC) is one of today’s hottest new technologies. However MCC goes beyond the mere access to data storage and processing services outside the mobile device, and despite its rapid growth, many issues still need to be addressed. In the third article “AContext-AwareArchitecture Supporting Service Availability in Mobile Cloud Computing”, Gabriel Guerrero-Contreras, Jos e Luis Garrido, Sara Balderas-D ıaz, and Carlos Rodr ıguez-Dom ınguez explore the use of autonomic computing techniques to cope with disconnections and network partitions which are inherent to the dynamic nature of mobile networks, andmay compromise service availability. In scenarios such as emergency situations and tourism, people may use local collaborative tools exploiting distributed resource pools provided by nearby mobile nodes. For such scenarios, the authors propose a reusable and adaptable service-oriented software architecture which uses a combination of service replication, self-configurable replicas’ activation/hibernation, contextawareness and election mechanisms, to satisfy availability requirements of mobile clouds. The proposal has been evaluated through the simulation of some scenarios representative of emergency situations, where traditional network infrastructures may be unavailable, and it is necessary to support the collaboration betweenwork groups through data storage and processing services. S. Russo is with the Dipartimento di Ingegneria Elettrica e Tecnologie dell’Informazione, Universit a di Napoli Federico II, Via Claudio 21, Napoli 80125, Italy. E-mail: [email protected]. M. Vieira is with the Departamento de Engenharia Inform atica, Universidad de Coimbra, Polo II Pinhal deMarrocos, Coimbra 3030-290, Portugal. E-mail: [email protected].